Using Rule Based Classifiers for the Predictive Analysis of Breast Cancer Recurrence

Abstract

The Aim of this work is to assess the Effectiveness of Rule Based Classifiers to help an Oncology Doctor for prediction of Breast Cancer Recurrence ,286 Cancer patient data ,obtained from UCI Machine learning Repository ,are used to determine Recurrence Events for New patients .This dataset is processed with WEKA Data Mining Tool ,by applying Rule Based Classifiers(RIPPER,DT,DTNB) and Rule Set is generated .Further from Experimental Results, it has been found that DTNB is providing improved Accuracy compared to other two Classifiers .Based on the patients’ characteristics and the Rule set generated by DTNB ,New patients may be labeled as developing or not Recurrence Events,thus supporting an Oncology Doctor in making Decisions about disease in a shorter time.

Keywords: Medical Data Mining, KDD, Breast Cancer Recurrence, Classification, Association, WEKA, Rule Based Classifiers